Identifying and counseling individuals at high risk of colon cancer because of their family history is complex and increasingly common. Prediction algorithms that fully exploit laboratory-based biological knowledge of inheritance mechanisms and other characteristics of susceptibility genes can contribute importantly to improved screening, prevention, and genetic testing, and to the design of health behavior interventions, cancer on studies, and genetic epidemiology studies. Mendelian risk prediction models make efficient use of laboratory-based information, and have been successfully used in a number of prevention, clinical, and research activities associated with the BRCA genes. The goals of this project are to develop (aim 1), and validate both retrospectively and prospectively (aim 2), a Mendelian risk prediction model for colon cancer susceptibility genes. The model will assist counseling of individuals at high risk of familial colon cancer, by providing accurate estimates of the probability that they carry a deleterious mutation of a colon cancer susceptibility gene, and of the probability that they subsequently develop cancer, based on family history. Syndromes considered will include HNPCC, FAP and the syndrome induced by the I1307K mutation of the FAP gene. Model training and validation will be based on a comprehensive elaboration of all evidence available, as well as data from four participating centers. Results from our study will make a significant contribution to colon cancer prevention by (a) promoting decision-making about genetic testing for susceptibility to colon cancer, via reliable individualized pre-test and post-test carrier probabilities; and (b) providing tools for identifying individuals at high risk of familial colon cancer. In addition, results from our study will also contribute to our understanding of genetic susceptibility to colorectal cancer by (a) developing reliable admission criteria for studies of gene characterization and gene environment interactions; (b) providing validated surrogate measures of genetic susceptibility for studies in which genetic testing is impractical or only partially performed.